Seeing Through the Tectal Eye: Visual Representations in the Primate Superior Colliculus With and Without Eye Movements

Abstract

Vision is an important sensory modality for primates. However, because of fo-veated retinal organization, vision requires repetitive eye movements to align the fovea with new objects. This creates interesting theoretical questions about percep-tion in general, since eye movements themselves alter images on the retina even if there are no moving objects in the world. Thus, to study vision is to also study how vision operates during active behavior. In my dissertation, I have investigated the concept of “active vision” in a brainstem structure critical for eye movement gener-ation, the superior colliculus (SC). The SC is a well-studied structure, with a promi-nent role in driving eye movements. However, this structure is also ultimately a vis-ual structure, and it is the primary visual structure in lower animals. Given a rela-tively sparse interest in visual properties of the primate SC in the literature, and given the proximity of both visual and motor representations already together within the same structure, we have adopted the SC as an ideal locus for investigating active vision. We first characterized SC visual representations in the absence of eye movements. We found surprising asymmetries in visual representations between upper and lower visual fields, which have direct consequences on oculomotor be-havior. We also performed analogs of visual neurophysiology experiments in struc-tures like primary visual cortex (V1) or lateral geniculate nucleus (LGN), but this time to characterize SC spatial and temporal frequency tuning properties. We found remarkable tuning properties and response time profiles of SC neurons that we think allow this structure to be highly in-tune with the statistics of natural scenes. This in turn allows very efficient eye movement response times to spatial frequencies prominent in our environment. In the same set of studies, we also characterized cen-ter-surround interactions, orientation tuning, and temporal frequency tuning. To fur-ther explore the concept of “active vision”, we showed how visual representations in the SC are modulated around the time of eye movements. We discovered surprising and spatially far-reaching pre-movement enhancement of contrast sensitivity, which can provide a neural basis for attentional enhancements in behavior. We also found spatial-frequency-specific post-movement modulations of neural activity. The latter results are particularly interesting when related to classic perceptual phenomena of saccadic suppression, and also when considering different neuronal cell types. Fi-nally, we tested how the eyes stabilize themselves after saccadic eye movements and found an enhanced ocular drift control even for the smallest possible saccades gen-erated during fixation. The overall aggregation of our results creates several inter-esting new research avenues with important and solid foundations for future under-standing of detailed circuit-mechanisms of SC function, and also for relating such mechanisms to perception and action.